BackgroundSulfamethoxazole-trimethoprim (SMX/TMP) is a standard drug for the prophylaxis of Pneumocystis pneumonia (PJP) in immunosuppressed patients with systemic rheumatic diseases, but is sometimes discontinued due to adverse events (AEs). The objective of this non-blinded, randomized, 52-week non-inferiority trial was to quest an effective chemoprophylaxis regimen for PJP with a low drug discontinuation rate. Results at week 24 were reported.MethodsAdult patients with systemic rheumatic diseases who started prednisolone ≥0.6 mg/kg/day were randomized into three dosage groups: a single-strength group (SS, SMX/TMP of 400/80 mg daily), half-strength group (HS, 200/40 mg daily), and escalation group (ES, started with 40/8 mg daily, increasing incrementally to 200/40 mg daily). The primary endpoint was non-incidence rates (non-IR) of PJP at week 24.ResultsOf 183 patients randomly allocated at a 1:1:1 ratio into the three groups, 58 patients in SS, 59 in HS, and 55 in ES started SMX/TMP. A total of 172 patients were included in the analysis. No cases of PJP were reported up to week 24. Estimated non-IR of PJP in patients who received daily SMX/TMP of 200/40 mg, either starting at this dose or increasing incrementally, was 96.8–100% using the exact confidence interval as a post-hoc analysis. The overall discontinuation rate was significantly lower with HS compared to SS (p = 0.007). The discontinuation rates due to AEs were significantly lower with HS (p = 0.006) and ES (p = 0.004) compared to SS. The IR of AEs requiring reduction in the dose of SMX/TMP (p = 0.009) and AEs of special interest (p = 0.003) were different among the three groups with significantly higher IR in SS compared to HS and ES.ConclusionsAlthough there were no PJP cases, the combined group of HS and ES had an excellent estimated non-IR of PJP and both were superior in safety to SS. From the perspective of feasibility and drug discontinuation rates, the daily half-strength regimen was suggested to be optimal for prophylaxis of PJP in patients with systemic rheumatic diseases.Trial registrationThe University Hospital Medical Information Network Clinical Trials Registry number is UMIN000007727, registered 10 April 2012.Electronic supplementary materialThe online version of this article (doi:10.1186/s13075-016-1206-8) contains supplementary material, which is available to authorized users.